Is there any mercury vapor ballast with a capacitor that can't run metal halide?

Is there any metal halide  probe start  ballast that can't run mercury vapor?

Why you write your questions in triple posting instead in one post?
1. CWA probe-start metal halide ballast needs a capacitor for proper lamp operation.
2. CWA and HX mercury ballasts have 240V OCV. This isn't sufficient for cold starting probe-start metal halide lamps. Probe-start metal halide lamp may be able to start from 240V but not reliablity, because of their higher ignition voltage than mercury lamps. To ensure that the lamp always starts during its entrie life, CWA ballast of probe-start metal halide lamps, have OCV of 380V. In 220-240V countries, it is possible to run probe-start metal halide lamps on a simple mercury choke and 750V two wires ignitor. It is also possible to use a fluorescent starter (Must be connected in series with a 60-100W incandescent lamp to limit current) parallel to the mercury ballast.
3.CWA ballast for probe-start metal halide lamps can run mercury lamps without any problem, since the lamp voltage and current of mercury lamps and probe-start metal halide lamps are the same with only the ignition voltage is different.

If sox lamps are more efficient than LEDs why do they use LEDs on the streets?

Because the DSA (dip-sheit association) wants whiter light.

Because the DSA (dip-sheit association) wants whiter light.

True

A bit more seriously:
SOX are physically large ight sources, so not possible to shape the beam as efficiently as with other HID's and mainly LED's (the efficiency is proportional to ratio of the optical system size vs light source size, while the usable lantern size is limited to about 50cm..1m).
So (for now assume the matching pattern was really used for a given task) the LED may suffice with 4000lm, but the LPS may need for the same task more than 20klm of light output. So even when the LPS may have 220lm/W, it would mean 90W SOX (neglecting the ballast losses), while for the LED lantern it would mean just about 50W for a common 80lm/W (at the end of rated life) LED. Plus the LED is today cheaper to make and gives off white light (so considered as way better light quality).
Of course, there are places, where the SOX have still perfect sense:
- Surrounding an observatory station (as the monochromatic light is way easier to filter out from the light "smog"), but with modern observatory placement (in the middle of nowhere on some high elevation site) that means just the lanterns lighting the site and these is even easier to just switch OFF for the shooting time
- In places with frequent fogs and requiring good visibility even at that condition. With the fog all the beam shaping looses it's point, as the fog diffuses it out again, so only the "brute force" with high total lumen output is working, plus the monochromatic and at the same time longer wavelength light of the LPS does not scatter that much, so offers better visibility at that condition.
But these two are rather special case (well, for someone living in a harbor city plus with frequent foggy weather it may seem normal), most cities around the globe do not suffer from these problems, so the LED's are there more appropriate (when judging about just the rated performance)

Of course, the LED is rather new technology, so the industry is just learning how to use it in really proper way (from LED die design and production, till street light system maintenance), so causing quite frequent errors anywhere in the complete chain, so making the LED's still not the cheapest "most universal" street light technology (that is still hold by the HPS; mainly when spot fixes are expensive)

What kind of light is this?
https://vine.co/v/eQEgw76Jd2w

Uhh your link is broken, Alex.
Here's the proper formatting for naming a link:
Name
Here it is in action:
Lighting Gallery

Uhh your link is broken, Alex.
Here's the proper formatting for naming a link:
Name .
Here it is in action:
Lighting Gallery

Fixed 

Looks like a cheap lithonia wraparound mounted in a crappy drop ceiling.

Medved
LED lanterns today may be directing more of the light to where it is wanted, but their optics (the lens panel over series of LED dots) does not provide any large emitting area, so the lantern itself - visible in the scene to the drivers, is basically a "laser in the eye"...

When you count this in, the same Lux on the road just dont matter anymore for actual vision, even when that and uniformity are the only thing standards demand...

What kind of light is this?
https://vine.co/v/eQEgw76Jd2w

Medved
LED lanterns today may be directing more of the light to where it is wanted, but their optics (the lens panel over series of LED dots) does not provide any large emitting area, so the lantern itself - visible in the scene to the drivers, is basically a "laser in the eye"...

When you count this in, the same Lux on the road just dont matter anymore for actual vision, even when that and uniformity are the only thing standards demand...

That problem is not special to LED's, that problem applies for all spot light sources, if you focus light from an equivalent MH, you will get the same lantern brightness (so same "laser", if you called it that way), it will just spill way more of the light and so require higher raw lumens.
Of course, it depends on the design of the given lantern. But mainly it depends on the pole distance-to-height ratio in the installation. Very often the poles are just too far away, so the lanterns are then adjusted to long throw to still cover the area between them. But that of course means more glare.

And there is one other aspect: When you compare an old, unfocused light installation with just illuminated islands around the poles vs the new one with the whole stretch illuminated, it makes you to focus your attention to just comparing the brightest spots. There of course the old appear "better", because the brightest spots are just brightest, with the light landing nearly vertically, so no glare at all.
But for the safety are most important the darkest spots (there the potholes in the road tend to hide the most) - and ironically these you will ignore at all.
It is not your fault, it is just how the subjective human nature works.
So you have to really evaluate the visibility across the whole area and take the worst spot as the final result...